The present invention relates to wireless communications, and more particularly to a dynamic threshold detector that enables more accurate detection of synchronization signals of transmitted packets to achieve more successful packet timing estimation.
In packet radio systems, transmitted packets typically consist of preamble and data payload. The wireless receiver uses the preamble to determine timing and process the data payload. Typically, timing estimation is achieved by correlating for a predetermined preamble sequence and detecting correlation via crossing of a threshold. The preamble sequences may be referred to as synchronization signals that are pre-pended to the packet prior to transmission. For an orthogonal frequency division multiplexing (OFDM) configuration according to the Institute of Electrical and Electronics Engineers, Inc. (IEEE) 802.11a standard, for example, a pair of predetermined long sync (LS) symbols are used as the preamble sequence. Several issues associated with threshold detection include threshold crossing at the correct time position, threshold crossing at an incorrect time position, and no threshold crossing.
In a multipath environment, preamble correlation is spread over several output samples so that the correlation peak is smaller. In general, the transmitted signal traverses multiple paths to the receiver resulting in an interference pattern that obscures the results of the correlation process. The correlation process operating in the multipath environment exhibiting a multipath delay profile results in multiple attenuated decaying peaks rather than one strong peak. A single, constant threshold value used to distinguish correlation peaks from noise is inadequate to achieve desired performance levels. A constant threshold value optimized for Additive White Gaussian Noise (AWGN) increases the occurrence of xe2x80x9cmiss eventsxe2x80x9d since the threshold must be set at a relatively high level that is greater than the more attenuated correlation events. A smaller threshold used for correlation detection catches the weaker signals but also increases the occurrence of false alarms. Furthermore, the characteristics of the wireless medium changes from one environment to another, so that the noise level and/or multipath delay profile may also vary. These factors may further change over time in a given environment, such as movement of interfering elements or periodic activation/deactivation of interfering communication devices. A single, constant threshold is inadequate to detect correlation peaks in the face of all of these potential contingencies thereby resulting in poor and/or variable performance levels.
The receiver typically performs automatic gain control (AGC) functions to control the power of received signals to a target power level to detect and acquire both strong and weak signals. AGC errors also significantly impact performance, particularly for radios employing a constant correlation detection threshold. AGC errors of positive or negative 3 decibels (dB) are typically expected. Negative AGC errors reduce the correlation peak and increase miss events. Positive AGC errors enhance noise level and increase false alarm events. The OFDM preamble has relatively high (poor) cross-correlation properties, which increases false alarm events. It is appreciated that the performance of the radio transceiver is very sensitive to the thresholding technique employed. It is desired to find a thresholding technique that enables relatively high performance levels regardless of environment, multipath and AGC discrepancies.
A threshold detector for a wireless receiver according to an embodiment of the present invention is configured to detect synchronization signals of packets transmitted in a wireless medium. The detector includes a correlator, short-term and long-term integrators, a multiplier and a detector. The correlator compares received signals with predetermined synchronization signals and provides a correlation signal indicative thereof The short-term integrator integrates a predetermined short term of the correlated signal and provides a short term moving average signal. The long-term integrator integrates a predetermined long term of the correlated signal and provides a long term moving average signal. The multiplier multiplies the long term moving average signal by a predetermined constant and provides a dynamic threshold signal. The detector asserts a timing signal based on a crossover between the short term moving average signal and the dynamic threshold signal.
The detector may include phase removal logic that removes phase of the correlation signal. The correlator may be configured to output a consecutive series of samples at a selected sample rate of the received signal as the correlation signal. In this case, the short-term integrator determines a moving average using a first predetermined number of samples based on a predetermined channel delay spread. The long-term integrator determines a moving average using a second predetermined number of samples selected to track noise level in the channel. The second predetermined number is greater than the first predetermined number so that the short term moving average signal is more responsive to correlation fluctuations.
The threshold detector may be used as a synchronization timing estimator in a baseband processor used for wireless packet-based communications, such as those used for WLAN applications. In an orthogonal frequency division multiplexing (OFDM) embodiment, for example, the threshold detector comprises timing estimation logic used to estimate the timing of long sync symbols in the preamble of OFDM packets.